Fig 2: Targeting EGFR/MEKs/ERKs reverses BETi resistance.a Measurement of cell viability of MDA-MB231 and BT549 cells by CellTiter-Glo® luminescent viability assay. Control and RPS6KA2-knockdown cells were treated with JQ1 in the absence or presence of seven EGFR inhibitors for 48 h. 1: Ctr; 2: osimertinib; 3: AZ5104; 4: erlotinib; 5: icotinib; 6: gefitinib; 7: lapatinib; 8: EAI045; 9: RPS6KA2-shRNA; 10: JQ1; 11: JQ1 + osimertinib; 12: JQ1 + AZ5104; 13: JQ1 + erlotinib; 14: JQ1 + icotinib; 15: JQ1 + gefitinib; 16: JQ1 + lapatinib; 17: JQ1 + EAI045; 18: JQ1 + RPS6KA2-shRNA. Statistical data (Mean ± SD) are shown. P values were calculated between single JQ1-treated samples and samples treated with JQ1 plus EGFR inhibitors. Symbol ‘§’ indicates statistical significance (P < 0.05, one-way ANOVA). b MDA-MB-231 cells were treated with JQ1 (1 µM) and/or osimertinib, GDC0994, XMD8-92 (1 µM) or GDC0994/XMD8-92 for 12 h, western blotting was performed to detect JunD protein expression, and phosphorylated and total levels of MEK1/2, ERK1/2, ERK5 and RSK3. c CCK8 assays were done to measure the killing effects. Control and RPS6KA2-knockdown MDA-MB-231 and BT549 cells were treated with indicated the inhibitors for 48 h. 1: Control; 2: osimertinib; 3: BIX-02188; 4: U0126; 5: U0126 + BIX-02188; 6: GDC0994; 7: XMD8-92; 8: GDC0994 + XMD8-92; 9: RPS6KA2-shRNA; 10: JQ1; 11: JQ1 + osimertinib; 12: JQ1 + BIX-02188; 13: JQ1 + U0126; 14: JQ1 + BIX-02188 + U0126; 15: JQ1 + GDC0994; 16: JQ1 + XMD8-92; 17: JQ1 + GDC0994 + XMD8-92; 18: JQ1 + RPS6KA2-shRNA; 19: JQ1 + RPS6KA2-shRNA + BIX-02188 + U0126; 20: JQ1 + RPS6KA2-shRNA + GDC0994 + XMD8-92; 21: JQ1 + RPS6KA2-shRNA + osimertinib. Statistical data (mean±SD) are shown. P values were calculated when compared between single JQ1-treated samples and samples that were treated with JQ1 plus kinase inhibitors. Symbol ‘§’ indicates statistical significance (P < 0.05, one-way ANOVA). d CCK8 assay was performed to observe inhibition effects. Vector control and RSK3 (upper) or JunD-overexpressing (bottom) MDA-MB-231 cells were treated with the indicated inhibitors for 48 h. 1: DMSO; 2: JQ1; 3: JQ1 + osimertinib; 4: JQ1 + U0126; 5: JQ1 + BIX-02188; 6: JQ1 + BIX-02188 + U0126; 7: JQ1 + GDC0994; 8: JQ1 + XMD8-92; 9: JQ1 + GDC0994 + XMD8-92. Statistical data (mean±SD) are shown. P values were calculated when compared between JunD or RSK3-overexpression clones and vector controls in the presence of the same inhibitors. Symbol ‘§’ indicates statistical significance (P < 0.05, one-way ANOVA), and symbol ‘#’ indicates no significance (P > 0.05). Source data are provided as a Source Data file.

Fig 3: JunD/RSK3 signalling correlates to BET inhibition sensitivity.a Correlation analyses of JUND and RPS6KA2 in four datasets of breast cancer patients from GEO. GSE76275 was separated into two sub-groups: triple-negative (TN) and non-triple-negative (non-TN); GSE43358 was separated into two sub-groups: triple-negative (TN) and non-triple-negative (non-TN); GSE42568 was separated into two sub-groups: ER-positive (ER+) and ER-negative (ER−); GSE31448 was separated into two sub-groups: luminal A (lumA) and luminal B (lumB). Pearson Coefficients of correlation and p values are shown. b Gene differential expression analyses of the mRNA levels of JUND and RPS6KA2 were conducted in two datasets that contained four subtypes of breast cancer patients, including BLBC, luminal A, luminal B and HER2+. c The mRNA expression statuses of JUND and RPS6KA2 were detected in a panel of breast cancer cell lines by quantitative PCR assay. d Protein expression levels of JunD and RSK3 were detected in a panel of breast cancer cell lines by western blotting. e Luminal and HER2+ breast cancer cell lines as well as BLBC cell lines were treated with DMSO or JQ1 (1 μM) for 48 h, cell growth was detected by CCK8 assay. Statistical data (mean ± SD) are shown. f RPS6KA2 and JUND genes were knocked down in two JQ1-resistant breast cancer cell lines, cell growth was measured by CCK8 assay. Statistical data (mean ± SD) are shown (***P < 0.001, one-way ANOVA). Source data are provided as a Source Data file.

Fig 4: JunD-dependent RPS6KA2 transcription mediates BETi resistance.a Western blotting was performed to detect JunD protein levels, MDA-MB-231 and BT549 cells were treated with DMSO or JQ1 (1 μM) for 0, 12 and 24 h. b Photograph depicted the potential JunD binding site in the enhancer region of RPS6KA2 gene. Wild-type and mutant RPS6KA2 gene enhancer luciferase plasmids are shown. c Luciferase assays were performed in MDA-MB-231 and BT549 cells transfected with shRNA for JunD and control, in the presence of DMSO or JQ1 (1 μM) for 6 h. Data are reported as mean ± SD. d Luciferase assays were performed in HEK293T cells transfected with wild-type or mutant RPS6KA2 enhancer luciferase construct with or without JUND co-transfection. HEK293T cells were treated with DMSO or JQ1 (1 μM) for 6 h. Data are reported as mean ± SD. e Chromatin immunoprecipitation (ChIP)-qPCR assay was executed in JUND-knockdown BLBC cells and their vector controls treated with DMSO or JQ1 (1 μM) for 6 h. ‘SP’ indicates specific primers of ChIP for RPS6KA2 gene enhancer and ‘NSP’ indicates non-specific primers that recognize the region downstream of the 3’ end of the gene. f Western blotting was performed to detect the expression levels of JunD and RSK3 in control and JUND-knockdown clones of MDA-MB-231 and BT549. g JUND or RPS6KA2-knockdown MDA-MB-231 and BT549 cells as well as their vector control cells were treated with DMSO or JQ1 (1 μM) for 48 h, and luminescent cell viability assays were done to detect the killing effects (**P < 0.01, one-way ANOVA). h Measured tumoursphere formation in JUND or RPS6KA2-knockdown MDA-MB-231 and BT549 cells as well as their vector controls. The cells were also treated with DMSO or JQ1 (1 μM). Statistical data of numbers of tumoursphere were shown (*P < 0.05, **P < 0.01; one-way ANOVA). i Rescued expression of RSK3 in control or JUND-knockdown BLBC cells which were treated with DMSO or JQ1 (1 μM) for 48 h. Cell viability was measured by CellTiter-Glo® luminescent viability assay (**P < 0.01, ***P < 0.001; one-way ANOVA). j Western blotting was done to examine the expression levels of JunD and RSK3 in control or JUND-overexpressing clones of MDA-MB-231 and BT549. k JUND-overexpressing MDA-MB-231 and BT549 cells as well as their vector control cells were treated with JQ1 (1 μM) for 48 h, cell viability was detected by CellTiter-Glo® luminescent viability assay. Statistical data (mean ± SD) are shown (**P < 0.01, ***P < 0.001; one-way ANOVA). Source data are provided as a Source Data file.

Fig 5: JQ1-resistant BLBC cells are sensitive to combined therapies.a RPS6KA2 gene enhancer luciferase reporter activity was detected in two resistant clones as well as their parental cells. The data are reported as mean ± SD. ‘P’ indicates parental BLBC cells, and ‘R’ indicates resistant clones. b Measurement of the protein levels of JunD and RSK3 as well as the phosphorylated forms of RSK3 in resistant clones and parental BLBC cells. c MiR-548d-3p levels were measured by RT-PCR in resistant clones and parental BLBC cells. d Sequential chromatin immunoprecipitation was performed in parental and JQ1-resistant BLBC cells. e MDA-MB-231 and BT549 JQ1-resistant clones were treated with the indicated inhibitors for 48 h, and the effects were detected by CCK8 assay. 1: DMSO; 2: osimertinib; 3: GDC0994; 4: XMD8-92; 5: GDC0994 + XMD8-92; 6: U0126; 7: BIX-02188; 8: BIX-02188 + U0126; 9: JQ1; 10: JQ1 + osimertinib; 11: JQ1 + GDC0994; 12: JQ1 + XMD8-92; 13: JQ1 + GDC0994 + XMD8-92; 14: JQ1 + U0126; 15: JQ1 + BIX-02188; 16: JQ1 + BIX-02188 + U0126. Statistical data (mean ± SD) are shown. P values were calculated compared between single JQ1-treated samples and samples that treated with JQ1 plus kinase inhibitors. Symbol ‘§’ indicates statistical significance (P < 0.05, one-way ANOVA). f Cell growth was detected by CCK8 assay in vector control and RPS6KA2-knockdown JQ1-resistant clones. Cells were treated as following, 1: DMSO; 2: JQ1; 3: JQ1 + osimertinib; 4: JQ1 + GDC0994 + XMD8-92; 5: JQ1 + BIX-02188 + U0126. Statistical data (mean ± SD) are shown. P values were calculated when compared between RPS6KA2-knockdown clones and vector controls in the presence of the same inhibitors. Symbol ‘§’ indicates statistical significance (P < 0.05, one-way ANOVA), and symbol ‘#’ indicates no significance (P > 0.05). g A proposed model illustrating the underlying mechanism of BET inhibition resistance and potential combination therapies. Source data are provided as a Source Data file.